Household change relation and phase identification method and system based on voltage sequence similarity

The invention discloses a user-transformer relation and phase identification method and system based on voltage sequence similarity. The method comprises the following steps: acquiring daily voltage data of each user in a transformer area to construct a voltage sequence; different types of feature points are obtained from the voltage sequence based on an improved APAA algorithm, different types of feature points are selected according to the number of various types of feature points to carry out dimension reduction reconstruction on the voltage sequence, and a voltage feature sequence of each user is obtained; and introducing a dynamic time warping distance, and clustering the voltage feature sequence of each user by using a density-based clustering algorithm to obtain a user change relation cluster. According to the method, after error disturbance caused by too small uncertain fluctuation is eliminated through an improved APAA algorithm, a low-dimensional vector capable of reflecting voltage sequence feature